Assured Guaranty (AGO) and Financial Security Assurance (FSA), which Assured acquired in July, have done well to avoid collateralized debt obligations backed by mortgage backed securities, which have crippled the remainder of the financial guaranty industry. The credit crisis has taken a toll on so many people that it is hard not to hope that Assured can create a success story out of an industry so severely hurt by the credit crisis. It would be especially attractive if the investment offered an attractive valuation, strong prospects for future revenue and income generation; and solid financials.
Unfortunately, Assured offers the opposite. The current marked capitalization plus the $574 million raised via an equity offering is over 6 times annualized new business production. The 20% premium to book value would be attractive if Assured’s loss reserves were conservative. Unfortunately, this is far from the case. There is a very real risk that loss payments on residential mortgage backed securities (RMBS) will exceed Statutory loss reserves by a factor of four or more, easily enough to wipe out the Statutory surplus of all Assured’s financial guaranty subsidiaries.
In addition, Assured’s capital base is extremely leveraged against municipal credits, many of which will face severe fiscal crises exacerbated by massive pension and benefit obligations.
This article focuses on RMBS, the more immediate of these risks. First, two transactions wrapped by Assured are evaluated, followed by a discussion of implications for Assured’s overall RMBS portfolio. Finally, the article offers a few concluding remarks about rating agencies and analysts.
TWO RMBS TRANSACTIONS
Assured faces likely losses on over $10 billion of securities from hundreds of transactions. Evaluating potential losses on all or most of these is a painstaking exercise, and would make for equally painful reading. However, a great deal of insight about Assured’s overall credit risk can be gleaned from just two transactions.
The first is a Countrywide home equity loan securitization, CWHEQ 2007-D, wrapped 100% by Assured Guaranty Corp. (AGC). The second is a Harborview Mortgage negative amortization Alt-A option ARM first lien transaction, HVMLT 2006-12. Assured Guaranty Municipal Corp., (AGM, formerly known as FSA), wraps two tranches of this pool, 2A-1B and 2A-2C.
The CWHEQ transaction is simpler. It involves junior liens, so losses upon liquidation have averaged nearly 100%, and this will probably continue unless we see a robust rebound in home prices. In addition, there is no deductible (subordination), so Assured pays any losses not covered by the spread between the average interest rate on the loans, net of expenses, and the interest rates on the securities.
According to Bank of New York Mellon (BK) reports, the outstanding principal balance for CWHEQ 2007-D was $634 million as of November 2008. Since then, 28% ($176 million) has been paid down, and the balance now stands at $458 million. Of the 28%, about 12% was paid from the pool’s internal cash flows (HELOC principal and interest). The other 16% was paid by AGC.
New delinquencies have decreased from a peak in early 2009 of around 4.5% per month to an average of around 4% per month more recently. Most new delinquencies have turned into more serious delinquencies. The net result is that the percentage of delinquent HELOCs in the pool has increased from 12.5% a year ago to 16.0% as of November 2009.
If performance does not improve, future payments will exceed 60% of the outstanding principal. Assuming gradual improvement as bad loans are liquidated and an increasing proportion of borrowers begin to manage their debt loads, losses in the 30%-35% range (around $150 million) should be reasonably expected.
HVMLT 2006-12 is slightly more complicated. Wells Fargo (WFC) reports reveal that, as of November 2009, 56.6% of the $3.6 billion of the outstanding loans were at least one payment past due (including loans in the process of being foreclosed, acquired properties, and bankruptcies). For the past six months, on average around 4% of the borrowers have missed an initial payment each month, of which over three-quarters of these have subsequently missed a second payment (i.e., over 75% of 30-day delinquencies have rolled to 60-day delinquencies). Losses on liquidated loans have averaged around 55% of the loan balance prior to liquidation.
The tranches wrapped by AGM, 2A-1B and 2A-2C, are protected from losses by 11.9% subordination plus spread interest (the difference between the net interest from the mortgages and the interest paid on the securities, currently around 1% annualized). As of November, losses of 30% (13.1% in excess of 11.9% subordination) would wipe out 75% of class 2A-1B and 100% of 2A-2C.
42.6% (just over $1.5 billion) of the loans are over six months delinquent or worse. If 60% of these severely delinquent loans are liquidated at an average loss of 55% over the next 12-18 months, the 11.9% subordination will be wiped out. Even if no more loans go bad, existing delinquencies might be enough to burn through 2A-1B and 2A-2C.
The nature of these loans makes more defaults likely. For most of the performing loans in the pool, minimum payments are being made, so the principal balance is growing. Borrowers who make minimum payments until their loans begin to amortize normally face future accelerated payments on a growing mountain of debt.
Assuming losses burn through 2A-1B and 2A-2C, around 25% of the principal will probably have been repaid before they are erased. AGM will be responsible for the remaining 75% (around $275 million).
Interestingly, Assured’s web site shows BB ratings for these securities. Moody’s rates them Ca (excluding the AGM wrap). S&P does not publish ratings that ignore the insurance, but assigns CCC ratings to more senior securities, which only default if 2A-1B and 2A-2C are completely wiped out.
IMPLICATIONS FOR ASSURED’S RMBS PORTFOLIO
If CWHEQ 2007-D experiences $150 million of losses and the HVMLT tranches experience $275 million, it is likely that Assured will need to pay over $5.5 billion of RMBS losses (excluding resecuritizations). It would be laborious to outline losses on the entire RMBS portfolio, but it is not difficult to compare these two securities with the remainder of the below investment grade (BB or lower) RMBS that are shown in the list of structured finance transactions on Assured’s web-site.
Loan performance on CWHEQ 2007-D is reasonably representative of AGC’s below investment grade HELOCs. Therefore, if CWHEQ 2007-D experiences losses in the 30%-35% range, AGC’s other HELOCs should also see losses in this range. AGM’s HELOCs are a little worse, so losses should be closer to 40%. AGC’s and AGM’s Alt-A closed-end second lien transactions look worse, so losses should be even higher (50%+).
Delinquencies are higher for HVMLT 2006-12 than AGM’s other below investment grade first lien Alt-A and option ARM securities (56.6% vs. an overall average of around 45%), but subordination is lower. Some of AGM’s insured Alt-A securities are not junior to other securities, so losses will be lower on these. The overall average loss should be close to 45%.
Delinquencies on AGC’s Alt-A securities are similar to AGM’s, but subordination levels are a little higher, and fewer of the securities are junior to other securities, so losses should be closer to 35%. Losses on AGC’s below investment grade subprime first lien wraps should also be around 35%, assuming losses for CWHEQ 2007-D and HVMLT 2006-12 are near the percentages discussed above. Delinquencies and losses are higher, but most or all of these share losses pro rata with other senior securities once the subordinate tranches are eliminated by losses (i.e., they do not absorb losses for more senior securities).
Future loan performance is inherently difficult to predict, and the current situation is unprecedented, so losses could be much higher or lower than the numbers discussed above, but the Statutory loss reserves shown on page 5 of Assured’s third quarter operating supplement (here - pdf warning) are only realistic if delinquencies drop precipitously and average losses on defaulted loans improve. Moreover, the exposures are highly correlated. For example, if HVMLT 2006-12 generates large losses, Assured will almost definitely need to pay several billion dollars on other RMBS.
The discussion above does not apply to two other potential sources of additional RMBS related losses.
- Alt-A resecuritizations – The brief description of the Synthetic Alt-A option ARM portfolio listed on page 12 of Assured’s structured finance transaction list suggests that the underlying collateral may be mostly subordinate Alt-A tranches, in which case the transaction will probably be a near complete loss. In addition, the $3 billion on the six Private Residential Mortgage transactions listed on page 11 could produce billions of dollars of losses if the collateral is mainly senior support tranches that absorb losses for super senior securities. If the collateral is entirely super senior, there might be no losses. It would make sense for investors to demand breakdowns of the underlying collateral by vintage, average tranche subordination and thickness, and delinquency rates.
- AGC wraps two life insurance securitizations, Ballantyne Re and Orkney Re, with $900 million of notional collateral. These transactions are in default and a large proportion of the supporting collateral is Alt-A RMBS, but Assured provides virtually no details that allow investors to assess their loss potential.
Investors who believe that recent RMBS loan performance will not improve should short Assured aggressively. In this case, Assured will be obligated to pay over $7.5 billion in RMBS losses. Barring a massive capital injection, this will easily wipe out the combined Statutory capital of Assured’s financial guaranty subsidiaries (shown on page 5 of the operating supplement), leaving shareholders with little or nothing.
Institutional investors who think $150 million on CWHEQ 2007-D and $275 million on HVMLT 2006-12 are realistic loss outcomes would do well to short Assured while purchasing a few RMBS securities carefully selected to correlate with Assured’s portfolio. This strategy could earn investors annual returns of 15%-20% on the RMBS, and a windfall on the short position as losses force further downgrades and Assured goes into run-off.
This trade would work because RMBS market valuations are completely disconnected from Assured’s loss reserves. The fair value discount on Assured’s riskiest $20-$25 billion of RMBS ($16.8 billion rated below investment grade plus the riskiest investment grade securities) is in the 50%-60% range, or in the rough ball park of $12 billion. (The $12 billion is higher than the net derivative liability on Assured’s balance sheet partly because the derivative liability has been reduced by $6.7 billion (disclosed on page 40 of the 10-Q) for Assured’s nonperformance risk, and partly because it excludes financial guaranty policies, which would add another $6 billion if written as credit derivatives.
Assured would make an excellent investment for retail investors who believe Assured’s loss reserves, and are convinced that municipal defaults will remain very low. If this happens, the share price will probably double or triple over the next couple of years.
Institutional investors who share this optimistic view and are able to purchase distressed assets could also double or triple their investment (or better if leverage is used), but the downside risk is much less than Assured’s shares at current prices. A break even result on the RMBS investment would correspond to at least $10 billion dollars of losses for Assured, leaving shareholders with little or nothing.
Assured might have been a better investment if the company had chosen to continue doubling down on RMBS. This is effectively what the FSA acquisition did. Instead, it looks like management has committed to the path that is easier to sell to investors. Investors who bought the story might be disappointed to learn that Assured’s market share of new public finance issues has dropped by 50% since the end of the third quarter.
RATING AGENCIES AND ANALYSTS
Although it is not important to Assured as an investment, a few closing comments are offered about evaluations (or perhaps lack thereof) of Assured by rating agencies and equity analysts. The parties are different, but the behavior seems nearly as shameful as the oversight lapses discussed in a post from six months ago about MBIA (see here).
- Other than fraud or gross negligence, it is difficult to fathom a logical explanation for S&P’s AAA rating of any Assured subsidiary. The absurdity of S&P’s AAA financial guarantor ratings were discussed in a previous article. The AAA rating now seems even more absurd in light of AGC’s $179 million of Statutory surplus as of the end of the third quarter (shown in page 5 of the third quarter financial supplement), an amount that would be wiped out by 10 basis points of additional losses on insured credits. If retail municipal investors who relied on S&P’s ratings lose any money because municipal bond values drop as a result of Assured’s ratings being cut by S&P after RMBS loan performance does not stage a miraculous recovery, S&P should be responsible for their losses.
- Fitch’s ratings for Assured subsidiaries are equally ridiculous because they are completely inconsistent with Fitch’s RMBS loss projections. Fitch does not rate many of the securities in Assured’s portfolio, but assumptions similar to those projected by Fitch on comparable transactions imply in the ball park of $7.5 billion of RMBS losses for Assured.
- Perhaps the recent barrage of analyst upgrades of Assured is based on in-depth analysis, but it seems possible that they may have ignored the negatives. Did these analysts evaluate wrapped RMBS to see the high loss potential? Did they notice that AGC’s Statutory surplus is perilously close to 0, and there is a risk that Statutory capital for one of Assured’s subsidiaries could soon be wiped out after losses burn through the capital from Assured’s recent equity offering? Did they take the time to review Bond Buyer statistics showing that Assured’s dollar share of new public finance issues has dropped from over 9% in the third quarter to less than.4.5% in November?
Assured might also be criticized for not disclosing the drop in U.S. public finance market share or the likelihood of additional RMBS losses (assuming management considers them likely). However, if Assured has done nothing illegal, it is difficult to fault a company that appears to be fighting for its future as a living business.
Disclosure: Short AGO, have been long in the past